Abstract:We propose a novel class of risk-aware control barrier functions (RA-CBFs) for the control of stochastic safety-critical systems. Leveraging a result from the stochastic level-crossing literature, we deviate from the martingale theory that is currently used in stochastic CBF techniques and prove that a RA-CBF based control synthesis confers a tighter upper bound on the probability of the system becoming unsafe within a finite time interval than existing approaches. We highlight the advantages of our proposed approach over the state-of-the-art via a comparative study on an mobile-robot example, and further demonstrate its viability on an autonomous vehicle highway merging problem in dense traffic.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the safety issue under uncertainty, especially for the control of stochastic safety - critical systems. Specifically, the paper proposes a new class of Risk - Aware Control Barrier Functions (RA - CBFs) to control systems in stochastic processes, so that there is a tighter upper bound on the probability of becoming an unsafe state within a finite time interval. Existing methods are usually based on martingale theory, but these methods may be too conservative in practical applications, especially when the initial conditions have a serious impact on the probability of safety. ### Main Contributions: 1. **Introducing a New Class of Risk - Aware Control Barrier Functions (RA - CBFs)**: - Using the generator conditions in the stochastic level - crossing literature, an upper bound is derived to estimate the probability that the system becomes unsafe within a finite time interval. 2. **Proving the Advantages of the RA - CBF Controller**: - Through mathematical derivations and numerical studies, it is proved that the RA - CBF controller is more aggressive (i.e., less conservative) while ensuring a smaller upper bound on system risk. 3. **Verification in Practical Applications**: - In the highway merging problem of mobile robots and autonomous vehicles, the effectiveness of the RA - CBF controller is demonstrated. Especially in the highway merging problem, a 99% safety probability is required, and the experimental results show that the RA - CBF controller can successfully merge in dense traffic. ### Specific Problem Description: - **Background**: Traditional Control Barrier Functions (CBFs) have been successful in deterministic systems, but in stochastic systems, existing methods (such as stochastic CBFs based on martingale theory) may be too conservative and unable to provide effective safety guarantees. - **Objective**: Design a new control framework that can provide a tighter upper bound on risk in stochastic systems, while reducing conservatism and improving system performance. ### Solution: - **Design of RA - CBF**: By introducing a new CBF condition, the generator condition is used to derive the upper bound on system risk. - **Theoretical Analysis**: It is proved that under certain conditions, the RA - CBF controller can provide a smaller upper bound on risk than existing methods. - **Experimental Verification**: Through the highway merging problem of mobile robots and autonomous vehicles, the practical effect of the RA - CBF controller is verified. ### Conclusion: The paper proposes a new Risk - Aware Control Barrier Function (RA - CBFs), which can provide a tighter upper bound on risk in stochastic safety - critical systems and shows better performance and lower conservatism in practical applications. Future work will consider measurement noise and explore applications in predictive control.

Safety Under Uncertainty: Tight Bounds with Risk-Aware Control Barrier Functions

Safe Control for Nonlinear Systems With Stochastic Uncertainty via Risk Control Barrier Functions

Robust Control Barrier Functions with Uncertainty Estimation

Robust Control Barrier Functions using Uncertainty Estimation with Application to Mobile Robots

Safe Control Design through Risk-Tunable Control Barrier Functions

Bounding Stochastic Safety: Leveraging Freedman's Inequality with Discrete-Time Control Barrier Functions

A Risk-Aware Control: Integrating Worst-Case CVaR with Control Barrier Function

Learning-Based Control Barrier Function with Provably Safe Guarantees: Reducing Conservatism with Heading-Aware Safety Margin

Robust Control Barrier Function Design for High Relative Degree Systems: Application to Unknown Moving Obstacle Collision Avoidance

Advances in the Theory of Control Barrier Functions: Addressing Practical Challenges in Safe Control Synthesis for Autonomous and Robotic Systems

Safe Control in the Presence of Stochastic Uncertainties

Probabilistically safe controllers based on control barrier functions and scenario model predictive control

Safe Multi-Agent Interaction through Robust Control Barrier Functions with Learned Uncertainties

Safe robust adaptive control under both parametric and nonparametric uncertainty

Learning-based Parameterized Barrier Function for Safety-Critical Control of Unknown Systems

Parameterized Barrier Functions to Guarantee Safety under Uncertainty

Safe Control Synthesis Using Environmentally Robust Control Barrier Functions

Robustness of Control Barrier Functions for Safety Critical Control

Reinforcement Learning-based Receding Horizon Control using Adaptive Control Barrier Functions for Safety-Critical Systems

Safety‐critical Control for Robotic Systems with Uncertain Model Via Control Barrier Function